key: cord-0957549-ziv3lmnu
authors: Mills, M. G.; Bruce, E. A.; Huang, M.-L.; Crothers, J. W.; Hyrien, O.; Oura, C. A. L.; Blake, L.; Brown, A.; Hester, S.; Wehmas, L.; Mari, B.; Barby, P.; Lacoux, C.; Fassy, J.; Vial, P.; Vial, C.; Martinez, J. R.; Oladipo, O. O.; Inuwa, B.; Shittu, I.; Meseko, C. A.; Chammas, R.; Ferreira Santos, C.; Dionisio, T. J.; Garbieri, T. F.; Parisi, V. A.; Mendes-Correa, M. C.; dePaula, A. V.; Romano, C. M.; Bentim Goes, L. G.; Minoprio, P.; Campos, A. C.; Cunha, M. P.; Vilela, A. P. P.; Nyirenda, T.; Sawasawa Mkakosya, R.; Muula, A. S.; Dumm, R. E.; Harris, R. M.; Mitchell, C. A.; Pettit, S.; Botten,
title: An international, inter-laboratory ring trial confirms the feasibility of an open source, extraction-less "direct" RT-qPCR method for reliable detection of SARS-CoV-2 RNA in clinical samples
date: 2021-04-14
journal: medRxiv : the preprint server for health sciences
DOI: 10.1101/2021.04.10.21254091
sha: eac368119c71b684bd4715459d406bc8b8f6c831
doc_id: 957549
cord_uid: ziv3lmnu

RT-qPCR is used world-wide to test and trace the spread of SARS-CoV-2. Extraction-less or direct RT-PCR is an open-access qualitative method for SARS-CoV-2 detection from nasopharyngeal (NP) or oral pharyngeal (OP) samples with the potential to generate actionable data more quickly, at a lower cost, and with fewer experimental resources than full RT-qPCR. This study engaged ten global testing sites, including laboratories currently experiencing testing limitations due to reagent or equipment shortages, in an international inter-laboratory ring trial. Participating labs were provided a common protocol, common reagents, aliquots of identical pooled clinical samples and purified nucleic acids, and used their existing in-house equipment. We observed 100% concordance across labs in the correct identification of all positive and negative samples, with highly similar Ct values observed. The test also performed well when applied to locally collected patient NP samples, provided the viral transport media did not contain charcoal or guanidine, both of which appeared to potently inhibit the RT-PCR reaction. Our results suggest that open access, direct RT-PCR assays are a feasible option for more efficient COVID-19 testing as demanded by the continuing pandemic.

The global coronavirus disease (COVID-19) pandemic response depends on effective rollout of 77 recently approved vaccines and the use of nonpharmaceutical interventions to slow the spread of the 78 disease. Physical distancing supported by test-and-trace informed containment strategies has been 79 promoted worldwide [1] . The effectiveness of testing as a containment strategy requires the 80 implementation of accessible, affordable, reliable, and rapidly executable test methods that can meet the 81 rapid pace of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) transmission [2] [3] [4] . At 82 present, this goal remains largely unmet. 83

The majority of regional and national health laboratories around the world rely on reverse 84 transcription-quantitative polymerase chain reaction (RT-qPCR) SARS-CoV-2 virologic testing methods 85 such as those developed by the World Health Organization and the U.S. Centers for Disease Control and 86

Prevention (CDC) to support their public health programs [5, 6] . The methods themselves are robust and 87 have proven to be useful standards for detection and reporting. However, sample processing time and a 88 lack of supplies to support extraction as required to run this type of assay have resulted in widely reported 89 backlogs and shortages in the United States and around the world [7] . In regions that also suffer from 90 systemic financial and logistical challenges (e.g., Africa, the Caribbean, and South America), these 91 hurdles will continue to consistently impair reliable procurement of consumables, support for staffing, 92 and thus testing viability [8, 9] . Although the diversity and efficiency of commercial virologic and 93 serologic test methods expands weekly, most public health laboratories lack the resources (human and 94 capital) or remit to pivot to novel commercial methods. 95

To address these challenges, the nonprofit Health and Environmental Science Institute (HESI) 96

convened an international network of public and academic COVID-19 testing laboratories -the 97 Propagate Network -with the goal of collectively evaluating and disseminating practical, efficient, and 98 impactful open-access methods for SARS-CoV-2 detection. The Propagate Network and others have 99 identified "extraction-less" or "direct" real time-reverse transcription polymerase chain reaction (RT-100 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.10.21254091 doi: medRxiv preprint PCR) as an open-access qualitative method for SARS-CoV-2 detection from nasopharyngeal samples 101 with the potential to generate actionable data more quickly, at a lower cost, and with fewer experimental 102 resources than full RT-qPCR [10, 11] . The method allows for detection of SARS-CoV-2 viral ribonucleic 103 acid (RNA) with the omission of the most labor-intensive step -the RNA extraction step -and its 104 associated extraction reagents. Published intralaboratory studies indicate that the technique is internally 105 reproducible (with some loss of sensitivity compared to standard RT-PCR) and is effective in detecting 106 both true negatives and positives. Notably, direct RT-qPCR remains sufficiently sensitive to detect viral 107 RNA from patients most likely to be infectious (cycle threshold [Ct] < 33) [12] [13] [14] [15] . 108

The major goal of this Propagate Network study was to determine the practical utility of an open-109 access, direct RT-PCR assay [10] via an international, interlaboratory ring trial. The study engaged 10 110 global sites, including laboratories currently experiencing many of the testing limitations described above, 111

in a series of studies involving a common protocol, common reagents, aliquots of identical pooled clinical 112 samples, and purified nucleic acids, using their existing in-house equipment. Our results suggest that 113 open-access, direct RT-PCR assays are a feasible option for more efficient COVID-19 testing as 114 demanded by the growing pandemic. 115

The Propagate Network study was coordinated and partially funded by the international nonprofit 118 HESI as part of its global public health mission and via voluntary contributions of time and effort from 119 the participating partners. Special acknowledgment is given to the University of Washington Virology 120 Laboratory (UWVL) for their efforts to prepare and ship samples for this study and to the University of 121 Vermont Larner School of Medicine for their support in refining the study protocols and recruiting 122 partner laboratories. 123

Ten laboratories were recruited to participate in the trial for the detection of SARS-CoV-2 RNA from 124 patient nasopharyngeal swabs without RNA extraction using kits provided by UWVL ( Table 1) . 125

Laboratories participated voluntarily and were not offered any compensation for their participation. Due 126 to logistical shipping challenges, which were in large part brought on by the pandemic, samples were 127 unable to be sent to Malawi or Nigeria, underlying the hardships some areas face when testing relies on 128 reagents or materials from other countries. 129 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The samples generated and disseminated as part of Projects A and B were approved under a waiver of 134 consent by the University of Washington institutional review board (IRB; STUDY00000408). The de-135 identified samples were determined to be exempt because they were not considered human subjects manuscript. The committee made the statement "It is considered that, given the modality of the 166 study, where the identity is duly protected, and, given the importance from the public health point 167 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. samples. Project C characterized the feasibility of the direct method as applied to locally sourced samples 177 collected as part of regional public health testing efforts (Fig 1) . 178 (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.10.21254091 doi: medRxiv preprint Project A. To confirm that all reagents arrived safely and that every laboratory could perform the 184 direct RT-PCR method, each laboratory tested a set of eight nucleic acid samples purified from patients 185 with COVID-19 and supplied by UWVL, which included six blinded samples (three positive and three 186 negative), one sample identified positive, and one sample identified negative to serve as controls, plus a 187 laboratory-supplied no-template water control. Each laboratory reviewed the results of Project A with the 188 study coordinator to confirm that they had correctly identified 100% of the positive and negative blinded 189 samples before proceeding. 

206 Identical kits were prepared by UWVL and sent on dry ice to each participating laboratory. These kits 207 included reagents for the three possible projects (Fig 1) : blinded purified total nucleic acid samples plus 208 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. with water controls on each plate. Reactions to measure the SARS-CoV-2 N-gene (using CDC N2 236 primers and FAM-labeled probe) and human RNase P gene (using CDC RP primers and FAM-labeled 237 probe) were carried out for each sample in parallel. 238

For each sample, a mean C t value was computed by averaging individual C t values from all 240 laboratories. A C t value residual (for a given laboratory and sample) was defined as the individual C t 241 value minus the associated mean C t value. For data visualization, individual C t values and residual C t 242 values were plotted against mean C t values. Assay specificity and sensitivity was evaluated using the 243 negative and positive blinded samples. 244

High-titer stocks of SARS-CoV-2 were treated at 95°C for 10 min. The stock virus had a titer of >10 6 247 focus forming units (FFUs) per milliliter. After heat treatment there was more than a 5-log drop, with no 248 detectable foci after 10 min at 95°C (Fig 2) . 249 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. 

All participating laboratories correctly identified 100% of the positive (n = 3) and negative (n = 3) 256 blinded samples sent for purposes of confirming that the samples arrived safely and that the laboratory 257 was able to run the direct RT-PCR method. 258

CoV-2 test are its sensitivity and specificity (simply put, the ability to accurately distinguish the presence 261 versus absence of the viral RNA) and the consistency of its performance across laboratories. As an initial 262 approach, assay specificity and sensitivity was evaluated using 5 known-negative and 25 known-positive 263 samples that were tested in a blinded fashion by the 10 laboratories. For the five negative samples, a total 264 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. Quantitative agreement between laboratories. In addition to providing a qualitative 275 determination of the presence versus absence of virus, RT-PCR tests for SARS-CoV-2 can provide 276 additional value by reporting their C t value, which serves as a proxy for the amount of viral RNA present. 277

We therefore investigated the C t values reported for the blinded positive samples tested by the 278 participating laboratories. In general, the agreement between C t values from different laboratories was 279 good, with tighter agreement at lower average C t (higher viral loads) than at higher average C t (lower 280 viral loads ; Fig 3) . 281 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. We then evaluated the overall quantitative performance of each individual laboratory against the 287 sample-specific average C t value as determined by all 10 laboratories. The C t value residual for a given 288 laboratory and sample was defined as the C t value for the corresponding laboratory and sample minus the 289 sample-specific average C t value; the narrower their distribution within a laboratory, the more consistent 290 the relationship of C t values from the laboratory with the average C t value from all laboratories. Residual 291 C t values had overall similar variability across samples and were minimally affected by the actual viral 292 load (Fig 4) . The residuals appeared to be centered around zero for most laboratories (Fig 5) , with the 293 exception of laboratories 3 and 7, for which residuals appeared systematically negative (indicative of C t 294 values consistently lower than average), and laboratory 9, in which residuals tended to be positive 295

(indicative of C t values consistently higher than average). There was no evidence that the assay had a 296 lower sensitivity in this particular laboratory. 297 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. 10), these studies demonstrated average losses of sensitivity of between 1.5 and 3.8 cycles in RP C t value 308 (Fig 6a; Table 3 ) and between 2.6 and 4.8 cycles in N2 C t value (Fig 6b; Table 3 ), compared with direct 309 RT-PCR. For RP, this resulted in no failure to detect any sample from any of the four laboratories. For 310 samples in which N2 was detectable by both direct and extraction RT-PCR, the difference in C t values 311 between the two methods did not correlate with the C t value obtained by either method (Fig 6c) . 312

However, a few samples (6 of 93) that were detected between C t values of 28 and 39 by extraction RT-313 PCR were undetectable by direct RT-PCR, while other samples in that range were still detectable by the 314 same laboratories (Fig 6c) . (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. Direct, extracted C t (mean ± SD) RP 2.5 ± 2.5 1.5 ± 1.9 3.1 ± 0.9 −0.3 ± 1.4 3.8 ± 2.2 2.5 ± 2.3 N2 4.3 ± 3.2 2.6 ± 2.0 2.6 ± 1.3 −1.6 ± 1.4 4.8 ± 2.3 3.4 ± 3.0 325

For laboratory 9, direct RT-PCR yielded lower Ct values for both RP and N2 than extraction RT-326 PCR, with an average difference of −0.3 cycles and −1.6 cycles, respectively ( Table 3 ). The reason for 327 this unexpected result is not clear, but it may have been related to the laboratory's observation that 328 samples became "highly viscous" after the heating step (an observation not reported in any of the other 329 participating laboratories). As with any method, the authors recommend internal validation of the 330 approach prior to clinical implementation. 331

For two laboratories, direct amplification of both N2 and RP was unsuccessful in all samples, 332 including for samples with low N2 C t values (high viral loads) as measured by extraction RT-PCR. These 333 samples were later determined to have been collected in transport media containing ingredients that were 334 inhibitory to PCR, including charcoal and guanidine (e.g., ManTacc UTM and Jiangsu Rongye 335

Technology LinkGen media were reported as incompatible). Information on the brand/type of viral 336 transport media was not available for all samples used in Project C (this information is often not reported 337 with swabbed samples as provided to analysis laboratories). However, the following media were 338 specifically identified as compatible with this method (Hardy viral transport media, saline, and phosphate-339 buffered saline). 340

Overall, the direct approach worked effectively to detect samples deemed positive by standard RT-341 qPCR when samples were collected in media lacking charcoal or guanine. 342

This study confirms that the direct RT-qPCR method, initially described by Bruce et al. [10] , has the 344 potential to meaningfully contribute to global efforts to detect and contain the COVID-19 pandemic. This 345 study provides evidence that the direct RT-qPCR method is an efficient, reliable, and achievable method 346 for detection of SARS-CoV-2. Although the reproducibility of the method has been reported in single-347 laboratory studies previously, this study is the first to demonstrate that a globally diverse set of 348 laboratories operating with different equipment, clinical sample collection and handling conditions, 349 resource limitations, and operating practices can successfully implement the method. and Nigeria were unable to receive or analyze the Project A/B sample kits. While this was unfortunate, it 359 is emblematic of the challenges that the African continent (among others) continues to face in receiving 360 needed laboratory supplies and the importance of resource-sensitive methods development efforts such as 361 these. 362

In Project C of this study, Propagate partner laboratories were encouraged to use their own extraction 363 methods and locally collected samples with RT-qPCR reagents supplied by UWVL to compare results 364 from the direct method versus standard extraction-based PCR. The majority of the participating partner 365 laboratories were successfully able to apply the method and reliably detect RT-PCR-positive samples. 366

The importance of "ground testing" new methods was made evident when some of the laboratories were 367 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.10.21254091 doi: medRxiv preprint unable to detect any signal (N2 or RP) following the direct method despite using high-titer positive 368 samples as detected by standard methods. Laboratories experiencing this problem in some cases were 369 working with samples collected in commercial viral transport media that were often found to contain 370 charcoal or in inactivating media such as those containing guanidine. We hypothesize that these are 371 inhibitory to the RT-PCR reaction in the absence of an extraction phase. Similar inhibitory outcomes have 372 been subsequently identified by other laboratories [11, 17] . In other cases, the constituents of media were 373 unknown, so we were not able to hypothesize why the direct method was incompatible. We recommend 374 that laboratories seeking to employ the direct method for SARS-CoV-2 detection should conduct a small 375 pilot run (comparing results from direct and full PCR analyses on the same samples) to ensure that sample 376 media are compatible with this method. This pilot should be replicated if/when sample collection methods 377 or media are changed. 378

The success of this ring trial is of critical importance given the growing calls for COVID-19 379 screening as a containment strategy. The growing pandemic requires that we supplement definitive 380 clinical testing with scalable screening strategies that generate efficient, reliable results that can readily 381 inform public health action (e.g., quarantine and isolation) [2]. Non-PCR immunoassay antigen screening 382 kits have decreased sensitivity as compared to standard PCR but are widely utilized depending on the 383 country's COVID-19 pandemic testing strategy [18] . As anticipated per previous studies, the direct 384 method as applied to SARS-CoV-2 results in some loss in sensitivity compared to standard PCR. One 385 primary explanation for this observation is that RNA extraction typically concentrates RNA present in the 386 clinical sample (by eluting the sample in a smaller volume. In addition, there is a low level of inhibition 387 seen in clinical NP samples loaded directly into an RT-PCR reaction, and the sensitivity of the approach 388 drops when more than 3 ul of patient sample is used [10]. However, this loss is of lower significance to 389 the method's potential value as a public health screening tool. The direct method succeeds in all of the 390 areas of greatest contemporary need: 1) it reliably detects samples with RNA levels correlating to the 391 presence of live virus (and thus most potential for infectivity), 2) it provides the potential to optimize 392 throughput and reduce costs/logistics for SARS-CoV-2 testing, 3) it is an open-access methodology with 393 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted April 14, 2021. ; https://doi.org/10.1101/2021.04.10.21254091 doi: medRxiv preprint no commercial barriers or de novo equipment investment hurdles, and 4) it can be readily adopted by 394 most current public health or clinical laboratories with experience handling infectious samples [14, 15] . 395

We believe that the direct RT-qPCR method for SARS-CoV-2 screening is ripe for adoption in 396 laboratories seeking to reduce turnaround time for processing samples, experiencing challenges in 397 accessing extraction reagents, seeking to decrease costs, and/or looking to reduce the use, handling, and 398 disposal of chemicals in their laboratory. We do not propose this method as a substitute for samples 399 requiring ultrasensitive detection. As with the adoption of any new method, appropriate validation must 400 be conducted by the host laboratory. As standard RNA extraction reagents for PCR can cost $5-$6 USD 401 per extraction and millions of these tests are performed each day around the world, the potential savings for routine diagnostic testing of SARS-CoV-2 (personal communication). In the 3 months following 411 implementation, >40,000 samples were tested using this workflow. The laboratory observed several 412 critical advantages with this approach, including dramatically reduced extraction reagent costs and a 413 halving of the average laboratory turnaround time, despite increasing test volumes. Further, the 414 independence from specialized extraction reagents for routine testing alleviated pressure on supply chains 415 to meet the increased demand. These same positive impacts on testing efficiency are expected to apply to 416 other laboratories that adopt the method. 417

While no current testing or screening method is optimal to all situations, the direct method should be 418 considered as a viable, fit-for-purpose resource to address the growing need for population monitoring 419 All rights reserved. No reuse allowed without permission.

(which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. 

World Health Organization. COVID-19 strategy update

Rethinking COVID-19 test sensitivity -a strategy for 451 containment

All In': a pragmatic framework 453 for COVID-19 testing and action on a global scale

World Health Organization. Laboratory testing for coronavirus disease 2019 (COVID-19) in 456 suspected human cases

Detection of 2019 novel 459 coronavirus (2019-nCoV) by real-time RT-PCR

2019-nCoV) real-462 time RT-PCR diagnostic panel for emergency use only: instructions for use

Supply shortages impacting COVID-19 and non-COVID testing 465

Test scarcity means Latin America is 'walking blindly' on virus 468

Lack of Covid-19 testing undermines Africa's 'success

Direct RT-qPCR 474 detection of SARS-CoV-2 RNA from patient nasopharyngeal swabs without an RNA extraction step

Massive 477 and rapid COVID-19 testing is feasible by extraction-free SARS-CoV-2 RT-PCR

CoV-2 Infections and transmission in a skilled nursing facility

Viral RNA load as 483 determined by cell culture as a management tool for discharge of SARS-CoV-2 patients from 484 infectious disease wards

Virological 491 assessment of hospitalized patients with COVID-2019

Kinetics and 494 isotype assessment of antibodies targeting the spike protein receptor-binding domain of severe acute 495 respiratory syndrome-coronavirus-2 in COVID-19 patients as a function of age, biological sex and 496 disease severity

The impact of viral transport media on PCR assay results for the detection of 498 nucleic acid from SARS-CoV-2

Evaluation of analytical performance of seven rapid 500 antigen detection kits for detection of SARS-CoV-2 virus

Versatile and flexible 503 microfluidic qPCR test for high-throughput SARS-CoV-2 and cellular response detection in 504 nasopharyngeal swab samples

No reuse allowed without permission. (which was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity